基于图像块复杂度和分段距离函数的人眼定位算法研究

摘要

针对有复杂背景的人脸中眼睛定位问题,本文采用肤色分割原理分离出背景信息,在人眼检测过程中采用寻找最大图像块复杂度的小块的方法,在人眼定位过程中采用计算分段距离函数的方法.本文精选了200张单人照片进行仿真实验,实验结果表明,上述解决方案可以从有干扰背景,人头侧向一方,甚至戴眼镜和帽子等饰物的图像中快速精确地识别出人脸和人眼.%For the human eye positioning with a complex background the paper uses skin color segmentation principle to separate background information. In the human eye detection process, it used the approach of finding the small block of maximum image block complexity. and used segmention distance function the process of human eye positioning. This paper selected 200 single photos to conduct simulation and experimental results showed that these solutions could quickly and accurately identify images of human faces and human eye in the images of head laeral side, or even wearing glasses and hats and other accessories from the interference background.